def generate_actor_result(test_opt, src_img_path):
imitator = Imitator(test_opt)
src_img_name = os.path.split(src_img_path)[-1][:-4]
test_opt.src_path = src_img_path
if test_opt.post_tune:
adaptive_personalize(test_opt, imitator, visualizer=None)
else:
imitator.personalize(test_opt.src_path, visualizer=None)
action_list_dict = {'dance': MIXAMO_DANCE_ACTION_IDX_LIST,
'base': MIXAMO_BASE_ACTION_IDX_LIST,
'acrobat': MIXAMO_ACROBAT_ACTION_IDX_LIST}
for action_type in ['dance', 'base', 'acrobat']:
for i in action_list_dict[action_type]:
if test_opt.output_dir:
pred_output_dir = os.path.join(test_opt.output_dir, 'mixamo_preds')
if os.path.exists(pred_output_dir):
os.system("rm -r %s" % pred_output_dir)
mkdir(pred_output_dir)
else:
pred_output_dir = None
print(pred_output_dir)
tgt_smpls = load_mixamo_smpl(i)
imitator.inference_by_smpls(tgt_smpls, cam_strategy='smooth', output_dir=pred_output_dir, visualizer=None)
save_dir = os.path.join(test_opt.output_dir, src_img_name, action_type)
mkdir(save_dir)
output_mp4_path = os.path.join(save_dir, 'mixamo_%.4d_%s.mp4' % (i, src_img_name))
img_path_list = sorted(glob.glob('%s/*.jpg' % pred_output_dir))
make_video(output_mp4_path, img_path_list, save_frames_dir=None, fps=30)
def generate_T_pose_novel_view_result(test_opt, src_img_path):
imitator = Imitator(test_opt)
src_img_name = os.path.split(src_img_path)[-1][:-4]
test_opt.src_path = src_img_path
if test_opt.post_tune:
adaptive_personalize(test_opt, imitator, visualizer=None)
else:
imitator.personalize(test_opt.src_path, visualizer=None)
if test_opt.output_dir:
pred_output_dir = os.path.join(test_opt.output_dir, 'T_novel_view_preds')
if os.path.exists(pred_output_dir):
os.system("rm -r %s" % pred_output_dir)
mkdir(pred_output_dir)
else:
pred_output_dir = None
print(pred_output_dir)
tgt_smpls = create_T_pose_novel_view_smpl()
imitator.inference_by_smpls(tgt_smpls, cam_strategy='smooth', output_dir=pred_output_dir, visualizer=None)
save_dir = os.path.join(test_opt.output_dir, src_img_name)
mkdir(save_dir)
output_mp4_path = os.path.join(save_dir, 'T_novel_view_%s.mp4' % src_img_name)
img_path_list = sorted(glob.glob('%s/*.jpg' % pred_output_dir))
make_video(output_mp4_path, img_path_list, save_frames_dir=None, fps=30)
# clean other left
clean(test_opt.output_dir)
def generate_orig_pose_novel_view_result(opt, src_path):
opt.src_path = src_path
# set imitator
viewer = Viewer(opt)
if opt.ip:
visualizer = VisdomVisualizer(env=opt.name, ip=opt.ip, port=opt.port)
else:
visualizer = None
if opt.post_tune:
adaptive_personalize(opt, viewer, visualizer)
viewer.personalize(opt.src_path, visualizer=visualizer)
print('\n\t\t\tPersonalization: completed...')
view_params = opt.view_params
params = parse_view_params(view_params)
length = 180
delta = 360 / length
logger = tqdm(range(length))
src_img_true_name = os.path.split(opt.src_path)[-1][:-4]
save_dir = os.path.join(opt.output_dir, src_img_true_name)
mkdir(os.path.join(save_dir, 'imgs'))
print('\n\t\t\tSynthesizing {} novel views'.format(length))
for i in logger:
params['R'][0] = 0
params['R'][1] = delta * i / 180.0 * np.pi
params['R'][2] = 0
preds = viewer.view(params['R'], params['t'], visualizer=None, name=str(i))
# pred_outs.append(preds)
save_img_name = '%s.%d.jpg' % (os.path.split(opt.src_path)[-1], delta * i)
cv2.imwrite('%s/imgs/%s' % (save_dir, save_img_name), tensor2cv2(preds))
"""
make video
"""
img_path_list = glob.glob("%s/imgs/*.jpg" % save_dir)
output_mp4_path = '%s/%s.mp4' % (save_dir, src_img_true_name)
make_video(output_mp4_path, img_path_list, save_frames_dir=None, fps=30)
clean(opt.output_dir)
clean(save_dir)
# restore weights
g.saver.restore(g.sess, opt.weight_path)
sample = opt.sample
noise_seed = opt.noise_seed
if conf.model == 'biggan':
graph_inputs = graph_util.graph_input(g, sample+1, seed=noise_seed,
category=opt.category,
trunc=opt.truncation)
filename = os.path.join(output_dir, '{}_seed{}_sample{}.mp4'
.format(opt.category, noise_seed, sample))
else:
graph_inputs = graph_util.graph_input(g, sample+1, seed=noise_seed)
filename = os.path.join(output_dir, 'seed{}_sample{}.mp4'
.format(noise_seed, sample))
if opt.filename:
# override default filename
filename = opt.filename
# get the appropriate sample from the inputs
for k,v in graph_inputs.items():
if isinstance(v, np.ndarray) and v.shape[0] == sample+1:
graph_inputs[k] = v[sample][None]
# create video
video.make_video(g, graph_inputs, filename, num_frames=opt.num_frames,
max_alpha=opt.max_alpha, min_alpha=opt.min_alpha,
channel=opt.channel)
if test_opt.post_tune:
adaptive_personalize(test_opt, imitator, visualizer)
imitator.personalize(test_opt.src_path, visualizer=visualizer)
print('\n\t\t\tPersonalization: completed...')
if test_opt.save_res:
pred_output_dir = mkdir(os.path.join(test_opt.output_dir, 'imitators'))
pred_output_dir = clear_dir(pred_output_dir)
else:
pred_output_dir = None
print('\n\t\t\tImitating `{}`'.format(test_opt.tgt_path))
tgt_paths = scan_tgt_paths(test_opt.tgt_path, itv=1)
imitator.inference(tgt_paths,
tgt_smpls=None,
cam_strategy='smooth',
output_dir=pred_output_dir,
visualizer=visualizer,
verbose=True)
src_img_name = os.path.split(test_opt.src_path)[-1]
if test_opt.save_res:
output_mp4_path = os.path.join(
test_opt.save_res, 'test_%.4d_%s.mp4' % ('1', src_img_name))
img_path_list = sorted(glob.glob('%s/*.jpg' % pred_output_dir))
make_video(output_mp4_path,
img_path_list,
save_frames_dir=None,
fps=30)
visualizer = None
# set imitator
imitator = Imitator(test_opt)
if test_opt.post_tune:
adaptive_personalize(test_opt, imitator, visualizer)
imitator.personalize(test_opt.src_path, visualizer=visualizer)
print('\n\t\t\tPersonalization: completed...')
if test_opt.save_res:
pred_output_dir = mkdir(os.path.join(test_opt.output_dir, 'imitators'))
pred_output_dir = clear_dir(pred_output_dir)
else:
pred_output_dir = None
print('\n\t\t\tImitating `{}`'.format(test_opt.tgt_path))
tgt_paths = scan_tgt_paths(test_opt.tgt_path, itv=1)
imitator.inference(tgt_paths,
tgt_smpls=None,
cam_strategy='smooth',
output_dir=pred_output_dir,
visualizer=visualizer,
verbose=True)
mp4_path = test_opt.output_dir + test_opt.output_dir.split(
'/')[-3] + '_' + test_opt.output_dir.split('/')[-2] + '.mp4'
imgs = sorted(glob.glob('%s/*.jpg' % pred_output_dir))
make_video(mp4_path, imgs, save_frames_dir=None, fps=30)